White Noise Based Synchronous Active DC and AC Impedance Estimation from Front-End Inverters

The paper deals with a White Noise based active methodology to be used for the simultaneous estimation of AC grid and DC-side impedance from a front-end inverter. The considered system consists of an inverter connected to the grid with a suitable AC filter, assuming that both AC and DC sides exhibit resistive-reactive behavior. The proposed method is described starting from selecting the main parameters of White Noise and the frequency limits in the impedance estimation. The stimuli sent as a disturbance to the grid are sampled and analyzed using the Fourier transform, and then the parametric impedance division is performed using two different methods. Two methods have been proposed to have redundancy in the calculation and thus to obtain greater certainty in the estimated values: the well-known non-linear Least Square and the newly conceived proposed method. Experimental tests are carried out to validate the proposed method using a Hardware-In-the-Loop real-time emulator. Finally, the results and their percentage errors are reported, highlighting the relationship between the magnitude of the injected disturbance, the received signal from the grid, and the estimation accuracy. The final considerations are based on the results achieved from the Power Density Spectra of both the AC and DC sides. Results show that the proposed method is computationally 10 times faster than Least Square, providing better accuracy in the estimated quantities.